1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

CCNA - Subnetting Question

Discussion in 'Service Provider' started by Crazydave1990, Jun 18, 2012.

  1. Crazydave1990

    Crazydave1990 Bit Poster

    19
    0
    9
    Hi guys!

    Just getting my teeth into subnets at the moment - pretty interesting stuff tbf, a little confusing, but I understand the basic concepts, 1s = network bits, 0's equal host bits etc.

    I was reading a book, and it stated that Class A address can only be given to 126 companies?

    Now, I'm just wondering, if you factor that, when you set-up a LAN, it assigns private IP addresses, which then go through a Default Gateway to access the internet, the NAT/PAT protocols translate your IP, into the Public IP on the other side of the router???

    So theoretically, couldn't I have as many Class A address as I liked, as long as the class A address was on a private LAN? (albeit, a very big private lan haha!)

    So, the 126 addresses that it mentioned, is that only for public addresses?
     
  2. BB88

    BB88 Kilobyte Poster Gold Member

    370
    11
    76
    There are only a possibility of 126 Network IDs, but a possibility of 16,277,214 hosts per Network ID.

    One of those 16 million hosts through NAT would translate to your IP.

    I may not of explained that fully/understand it, as I am only half-way through my Network+ :D
     
    Last edited: Jun 18, 2012
    Certifications: AS Computing, A+, Network+, 70-680, 70-410
    WIP: MCSA: Server 2012
  3. Crazydave1990

    Crazydave1990 Bit Poster

    19
    0
    9
    Ok, so, if I get what your saying, BT Broadband for example own x amount of class A address, which it then distributes to it's customers???

    then internally, those routers will be configured to run a class C network? could they, for example, run a class B network on there? could they change there internal IP address to whatever they please internally?
     
  4. BB88

    BB88 Kilobyte Poster Gold Member

    370
    11
    76
    I see no reason why you couldn't run a Class A, B or C network internally.

    From what I read from NAT, each IP Address is assigned a Port. The maximum number of ephemeral ports is 65536. I guess that means you can only use upwards from 1024 for the internal IP addresses before they are translated with NAT.

    A single listening port can accept more than one connection simultaneously.
     
    Last edited: Jun 18, 2012
    Certifications: AS Computing, A+, Network+, 70-680, 70-410
    WIP: MCSA: Server 2012
  5. gmbaxter

    gmbaxter Bit Poster

    12
    0
    12
    Certifications: BSc (Hons) Computer Network Technology, ACSP, ACTS
    WIP: MCSA (2012)
  6. soundian

    soundian Gigabyte Poster

    1,460
    71
    107
    I'm not sure if that's the case. I would imagine a more flexible way would be to give ISPs a subnet of the addresses. For example, provider A could have a subnet mask of 255.128.0.0 and take the 0-127 part, and provider B could take the 128-255 part. How many ISPs need 16 million addresses? What a waste if they only need 2 million, which a 'class B' address wouldn't provide.
    Classes don't mean too much nowadays anyway. Handy when you're starting out, especially since you need to know what a class is before the classless part of CIDR makes any sense, but very few production systems are classful nowadays.
     
    Certifications: A+, N+,MCDST,MCTS(680), MCP(270, 271, 272), ITILv3F, CCENT
    WIP: Knuckling down at my new job

Share This Page

Loading...